JP2024020004A - Organism automatic measuring apparatus, method, and program - Google Patents

Abstract

To provide an organism automatic measuring apparatus, a method, and a program capable of measuring organisms such as Nishikigoi with high accuracy.SOLUTION: An organism automatic measuring apparatus comprises: a measurement target part recognition unit 12 for recognizing a measurement target part of an organism in a photographed image obtained by capturing the organism to be measured; and an automatic measurement unit 13 for automatically measuring the organism based on the measurement target part recognized by the measurement target part recognition unit 12.SELECTED DRAWING: Figure 1

Description

The present invention relates to an automatic biological measurement device, method, and program for automatically measuring living things based on captured images of living things.

Traditionally, many Nishikigoi have been bred as ornamental fish, and their body lengths and other measurements are carried out in order to manage their breeding and record their growth.

Furthermore, the body length of Nishikigoi is also used when evaluating the value of Nishikigoi. For example, Patent Document 1 proposes a system for objectively evaluating Nishikigoi based on photographed images of Nishikigoi, and proposes using the body length of Nishikigoi during the evaluation.

Patent No. 7008957

However, it may be difficult to manually accurately measure the body length of a swimming Nishikigoi. Furthermore, there is a risk of damaging the Nishikigoi by coming into contact with them during measurement. Furthermore, errors may occur depending on the measurement method used by the measurer, and the objectivity of Nishikigoi evaluation may not be guaranteed.

In addition to the above-mentioned Nishikigoi, there are cases in which it is desired to accurately measure the body length of goldfish, the height of plants, etc. for ornamental fish such as goldfish and ornamental plants in order to manage their breeding and record their growth.

In view of the above circumstances, an object of the present invention is to provide an automatic biological measuring device, method, and program that can measure the length of living things with high precision without damaging living things such as Nishikigoi. .

The automatic biological measurement device of the present invention includes a measurement target part recognition section that recognizes the measurement target part of the living organism in a photographed image of the measurement target living organism, and a measurement target part recognized by the measurement target part recognition part. and an automatic measurement section that automatically measures living things.

The automatic biological measurement method of the present invention can recognize a measurement target part of a living organism in a photographed image of the living organism to be measured, and perform automatic measurement of the living creature based on the recognized measurement target part.

The automatic biological measurement program of the present invention includes a step of recognizing a measurement target part of a living organism in a photographed image of the living organism to be measured, and a step of automatically measuring the living organism based on the recognized measurement target part. can be executed.

According to the automatic biological measurement device, method, and program of the present invention, the measurement target part of the living organism in a photographed image of the living organism to be measured is recognized, and the biological measurement is automatically performed based on the recognized measurement target part. This makes it possible to measure living things such as Nishikigoi with high precision.

A block diagram showing a schematic configuration of a Nishikigoi automatic measurement and evaluation system using an embodiment of the automatic biological measurement device of the present invention. Diagram showing an example of a shooting frame Diagram showing other examples of shooting frames Diagram showing an example of a graduated color palette Diagram showing the sRGB values of the color chart shown in Figure 4 A side view of the graduated color palette shown in Figure 4 as seen from the arrow X direction and a side view from the arrow Y direction. A diagram showing an example of a straight line connecting the tip of the head of a Nishikigoi and the tip of the caudal fin. Diagram showing an example of a chess box Diagram to explain the body length measurement of Nishikigoi using a chess box Diagram showing an example of outline information of Nishikigoi A diagram showing an example of the frequency distribution of the colors of each pixel that makes up the image of a Nishikigoi Diagram showing an example of a radar chart that is the evaluation result of Nishikigoi Flowchart for explaining the process flow of the Nishikigoi automatic measurement and evaluation system using an embodiment of the automatic biological measurement device of the present invention

DESCRIPTION OF THE PREFERRED EMBODIMENTS An automatic Nishikigoi measurement and evaluation system using an embodiment of the automatic biological measurement device of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a Nishikigoi automatic measurement and evaluation system 1 of this embodiment.

The Nishikigoi automatic measurement and evaluation system 1 of this embodiment includes a Nishikigoi automatic measurement device 10 and a Nishikigoi evaluation device 20, as shown in FIG. The Nishikigoi automatic measurement and evaluation system 1 of the present embodiment photographs a Nishikigoi to obtain a photographed image using the Nishikigoi automatic measurement device 10, and performs automatic measurement of the Nishikigoi and color correction of the photographed image based on the photographed image. The Nishikigoi automatic measurement and evaluation system 1 then uses the Nishikigoi evaluation device 20 to objectively evaluate the Nishikigoi based on the automatic measurement results by the Nishikigoi automatic measurement device 10 and the color-corrected photographed images.

The Nishikigoi automatic measurement device 10 and the Nishikigoi evaluation device 20 are communicably connected, for example, through a communication line such as the Internet or a LAN (Local Area Network).

First, the Nishikigoi automatic measuring device 10 will be explained. The Nishikigoi automatic measuring device 10 of this embodiment photographs a Nishikigoi underwater to obtain a photographed image, and automatically measures the body length of the Nishikigoi based on the photographed image. Furthermore, the Nishikigoi automatic measurement device 10 performs color correction on the photographed image to obtain the true color of the Nishikigoi.

Specifically, the Nishikigoi automatic measurement device 10 includes a photographing section 11, a measurement target part recognition section 12, an automatic measurement section 13, a color correction section 14, and a display section 15.

The Nishikigoi automatic measuring device 10 is composed of a mobile portable terminal such as a tablet terminal or a smartphone, and includes a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. It is equipped with storage such as semiconductor memory and hard disks, and communication I/F (Interface).

Further, a Nishikigoi automatic measurement program including an embodiment of the biological automatic measurement program of the present invention is installed in the semiconductor memory or storage of the Nishikigoi automatic measurement device 10. When the CPU or GPU executes this Nishikigoi automatic measurement program, the above-described photographing section 11, measurement target part recognition section 12, automatic measurement section 13, color correction section 14, and display section 15 function. Note that the hardware configuration of the Nishikigoi automatic measuring device 10 is not limited to the configuration described above.

Furthermore, in this embodiment, the functions of the above-mentioned photographing section 11, measurement target part recognition section 12, automatic measurement section 13, color correction section 14, and display section 15 are all executed by the Nishikigoi automatic measurement program. The present invention is not limited to this, and some or all of the functions may be configured from hardware such as an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or other electric circuits.

Further, the Nishikigoi automatic measurement program may be installed on the Nishikigoi automatic measurement device 10 as in this embodiment, or may be installed on a cloud server connected to the Nishikigoi automatic measurement device 10 via an Internet line. A Nishikigoi automatic measurement program may also be used.

The photographing unit 11 has an image sensor such as a C-MOS (Complementary Metal Oxide Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and a photographing optical system, and uses these to photograph a Nishikigoi swimming in the water. Obtain the captured image.

Here, in this embodiment, the body length of the Nishikigoi is measured using the photographed image taken by the photographing unit 11, but when photographing the Nishikigoi with a mobile mobile terminal such as a tablet terminal or a smartphone, the angle of the mobile mobile terminal The body length of the Nishikigoi in the captured image changes depending on the Nishikigoi's posture.

Therefore, the photographing unit 11 of the present embodiment displays a photographing frame on the display unit 15 that guides the photographing image so that the mobile terminal is at an appropriate angle and the Nishikigoi is in an appropriate posture. Let it be. FIG. 2 is a diagram showing an example of the photographing frame 15a displayed on the display unit 15 of the Nishikigoi automatic measuring device 10. As shown in FIG. 2, the photographing frame 15a is a contour image representing the outer shape of the Nishikigoi that is the object of measurement. Although FIG. 2 shows the photographing frame 15a when photographing a Nishikigoi from the front (dorsal fin side), for example, when the Nishikigoi automatic measuring device 10 is rotated 90 degrees as shown in FIG. The display may be switched to the photographing frame 15b for photographing the Nishikigoi from the side (pectoral fin side).

Then, the photographing unit 11 recognizes whether or not the Nishikigoi to be measured falls within the photographing frame 15a shown in FIG. 2 or the photographing frame 15b shown in FIG. Photography is permitted only if the person is present.

Specifically, the photographing section 11 causes the display section 15 to display the photographing frames 15a and 15b, and also causes the display section 15 to display a preview image captured by the image sensor. The preview image is a video.

Furthermore, the photographing section 11 causes the display section 15 to display a photographing button 14c, as shown in FIGS. 2 and 3. When the shooting button 14c is tapped, the shutter is released and a photographed image is taken. However, if the Nishikigoi in the preview image does not fit within the shooting frames 15a and 15b, even when the shooting button 14c is tapped. Prohibits taking images without releasing the shutter. In this way, by allowing photography only when the Nishikigoi in the preview image is within the shooting frames 15a and 15b, automatic measurement of Nishikigoi and can be evaluated.

Note that shooting can be prohibited by not releasing the shutter even if the shooting button 14c is tapped as described above, by hiding the shooting button 14c itself, or by changing the color of the shooting button 14c when shooting. The color may be different from that when possible (when the Nishikigoi is within the shooting frame 15a, 15b).

Then, the user adjusts the direction and angle of the automatic Nishikigoi measuring device 10 so that the Nishikigoi in the preview image fits within the photographing frames 15a and 15b. Then, the photographing unit 11 releases the shutter and photographs a photographed image only when the Nishikigoi in the preview image fits within the photographing frames 15a and 15b and the photographing button 14c is tapped.

Note that even if the Nishikigoi in the preview image falls within the shooting frames 15a and 15b, the imaging unit 11 detects that even if the Nishikigoi in the preview image is within the shooting frames 15a and 15b, if the amount of deviation between the outline of the Nishikigoi and the shooting frames 15a and 15b is greater than a preset threshold, , photography may be prohibited. Note that the outline of the Nishikigoi in the preview image can be extracted using known image processing.

Further, the photographing unit 11 automatically releases the shutter and photographs a photographed image when the amount of deviation between the contour of the Nishikigoi in the preview image and the photographing frames 15a, 15b becomes less than a preset threshold. Good too.

Further, in this embodiment, when photographing a Nishikigoi by the photographing unit 11 as described above, the graduated color palette 30 is placed within the photographing range, and the graduated color palette 30 is photographed together with the Nishikigoi. The image of the graduated color palette 30 is included in the image.

FIG. 4 is a diagram showing an example of the above-mentioned graduated color palette 30. The graduated color palette 30 has a color chart 31 and scale sections 32a to 32d. The color chart 31 is a 4×6 array of 24 rectangular patterns 31a to which 24 standard colors are respectively assigned. FIG. 5 is a diagram showing the relationship between the number of each rectangular pattern 31a and the sRGB value of that number. In addition, in this embodiment, the color chart 31 corresponds to the color sample of this invention.

The scale portions 32a to 32d are arranged around the color chart 31, and are arranged on each of the four sides of the color chart 31. In the scale portions 32a to 32d, scale patterns 32e are arranged at predetermined intervals in the length direction. The length d1 of the scale pattern 32e and the length d2 of the interval 32f are each set to 5 mm (d1=d2=5 mm).

The length of the long side of the graduated color palette 30 is 120 mm, and the length of the scale portions 32a and 32b is 100 mm. Further, the length of the short side of the graduated color palette 30 is 80 mm, and the length of the scale portions 32c and 32d is 50 mm.

Further, the graduated color palette 30 is formed by overlapping a first sheet 33 having a color chart 31 and a second sheet 34 having scale portions 32a to 32d. 6A is a side view of the graduated color palette 30 shown in FIG. 4 viewed from the arrow X direction, and FIG. 6B is a side view viewed from the arrow Y direction.

The measurement target part recognition unit 12 recognizes the measurement target part of the Nishikigoi in the photographed image. Specifically, the measurement target part recognition unit 12 of this embodiment recognizes the head and caudal fin of a Nishikigoi as the measurement target parts. The measurement target part recognition unit 12 has a trained model for measurement target recognition obtained by machine learning images of the head and caudal fin of a Nishikigoi, and a model that is photographed by the photographing unit 11 for this trained model for measurement target recognition. By inputting a photographed image, the head and caudal fin of a Nishikigoi in the photographed image are recognized. Note that when recognizing the part of the Nishikigoi to be measured, it is preferable to use a monochrome image in which the area of the Nishikigoi included in the photographed image is converted to white, and the area around the Nishikigoi is converted to black. In this case, the above-mentioned monochrome image is also used when performing machine learning on the trained model for measuring object recognition.

Note that a known method can be used as a machine learning method, such as a deep neural network (DNN) and a convolutional neural network (CNN).

The method for recognizing the head and caudal fin of the Nishikigoi in the photographed image is not limited to the method using the above-mentioned trained model for recognizing the measurement object, but other known pattern recognition methods may be used.

The automatic measurement unit 13 automatically measures the Nishikigoi based on the measurement target part (head and caudal fin of the Nishikigoi) recognized by the measurement target part recognition unit 12. The automatic measurement unit 13 of this embodiment automatically measures the body length of the Nishikigoi based on the recognized head and caudal fin of the Nishikigoi.

Specifically, the automatic measurement unit 13 determines a straight line connecting the tip of the Nishikigoi's head and the tip of its caudal fin, as shown in FIG. For example, the longest straight line connecting pixels forming the outline of the head and pixels forming the outline of the caudal fin is determined.

The automatic measurement unit 13 also measures the scale parts 32a to 32d that are most parallel to the straight line connecting the tip of the Nishikigoi's head and the tip of the caudal fin, among the scale parts 32a to 32d of the graduated color palette 30 included in the photographed image. is determined, and its scale portions 32a to 32d are extracted. Here, it is assumed that the scale portion 32a has been extracted.

Then, the automatic measurement unit 13 determines the number of scale patterns 32e of the scale part 32a included in the straight line connecting the tip of the head of the Nishikigoi and the tip of the caudal fin, and multiplies the number by 5 mm to determine the number of scale patterns 32e of the Nishikigoi. Measure the body length of.

In this embodiment, since the scale pattern 32e is set to 5 mm, the measurement error of the body length of the Nishikigoi is ±5 mm, which is sufficient for the accuracy of the body length of the Nishikigoi used in the evaluation of the Nishikigoi to be described later. On the other hand, if the scale is in 1mm increments like a regular ruler, for example, if it is submerged in water, it may not be possible to read the scale properly due to water refraction or fluctuations in the water surface, resulting in inconveniences such as inability to measure. .

Furthermore, when capturing an image by submerging the color palette 30 with a scale at the bottom of an aquarium containing Nishikigoi, for example, the distance from the automatic Nishikigoi measuring device 10 to the Nishikigoi and the distance to the color palette 30 with a scale are different. The accuracy of measuring the body length of Nishikigoi may decrease. Therefore, in such a case, the measured body length of the Nishikigoi may be corrected by performing a proportional calculation, taking into account the distance between the Nishikigoi and the Nishikigoi automatic measuring device 10. The distance between the Nishikigoi and the Nishikigoi automatic measuring device 10 may be measured using, for example, LiDAR (light detection and ranging). In this embodiment, the body length of the Nishikigoi is measured using the graduated color palette 30 as described above, but the body length of the Nishikigoi may also be measured using only the LiDAR described above.

Furthermore, the angle between the straight line connecting the tip of the Nishikigoi's head and the tip of the caudal fin and the length direction of the scale portion 32a is large, which may reduce the accuracy of measuring the body length of the Nishikigoi. Therefore, the measured body length of the Nishikigoi is corrected by performing trigonometric calculations taking into consideration the angle formed by the straight line connecting the tip of the Nishikigoi's head and the tip of the caudal fin and the length direction of the scale section 32a. You may also do so.

In addition, the photographed images are compared to a trained model for body length measurement in which the relationship between the straight line connecting the tip of the Nishikigoi's head and the tip of the caudal fin and any one of the scale parts 32a to 32d and the actual body length of the Nishikigoi is machine learned. The actual body length of the Nishikigoi is determined by inputting the straight line connecting the tip of the Nishikigoi's head and the tip of the caudal fin, which were found from Good too.

The body length of the Nishikigoi measured by the automatic measurement unit 13 is output to the Nishikigoi evaluation device 20 and used when evaluating the Nishikigoi.

The color correction unit 14 corrects the colors of the photographed image based on the colors of the color chart 31 in the photographed image. The color of the Nishikigoi included in the photographed image may differ from the actual color of the Nishikigoi due to the presence of the Nishikigoi in the water, and may be affected by water or lighting conditions. be.

Therefore, in this embodiment, color correction is performed on the photographed image so that the color of the Nishikigoi included in the photographed image approaches the color of the real Nishikigoi. Specifically, the color correction unit 14 extracts each color from the image of the color chart 31 included in the captured image, and calculates the sRGB value of the color extracted from the image and the sRGB value of each color shown in FIG. are compared, and color correction is applied to the photographed images so that the differences are canceled. Thereby, the color of the Nishikigoi included in the photographed image can be brought closer to the color of the real Nishikigoi, and the accuracy of evaluating the color of the Nishikigoi, which will be described later, can be improved.

The photographed image color-corrected by the color correction unit 14 is output to the Nishikigoi evaluation device 20 and used when evaluating the Nishikigoi.

The display unit 15 is composed of a touch panel having a liquid crystal display, etc., and displays a preview image when photographing a Nishikigoi, a photographed image of the Nishikigoi, a color-corrected photographed image, and the like.

In the above embodiment, the color chart 31 is used to measure the body length of Nishikigoi, but another method for measuring the body length of Nishikigoi will be described. Another method for measuring the body length of Nishikigoi uses a rectangular chess board 35 as shown in FIG.

The chess board 35 shown in FIG. 8 is an A5 size sheet-like printed matter. One square 35a of the chess board 35 is a square with a length of 29 mm and a width of 29 mm, and the black and white squares 35a are arranged in an array of 6×4=24 squares.

Next, a method for measuring the body length of a Nishikigoi using the chess board 35 shown in FIG. 8 will be described in more detail.

First, the "distance from the water surface to the carp" and the "water depth" are manually set and input by the user.

Next, using a sensor such as the above-mentioned LiDAR (light detection and ranging) that the Nishikigoi automatic measuring device 10 has, the center of the chess board 35 installed at an arbitrary position near the Nishikigoi and the Nishikigoi automatic measuring device 10 are connected. Measure distance. Note that at this time, the chess board 35 does not need to be submerged in an aquarium in which the colored carp swims, but may be fixedly installed near the colored carp.

Then, using the photographing unit 11 of the Nishikigoi automatic measuring device 10, the Nishikigoi and the chess board 35 are photographed together to obtain a photographed image for measurement.

Next, the automatic measurement unit 13 recognizes the image of the chess board 35 included in the measurement photographed image, and detects four preset points in the image. Specifically, as the four preset points, for example, as shown by arrows in FIG. 8, a point P where the vertices of four vertically and horizontally adjacent squares touch is detected.

Then, the automatic measurement unit 13 performs projective transformation processing on the measurement photographic image so that the aspect ratio of the square formed by the four detected points P becomes the correct aspect ratio (the aspect ratio of the actual object).

Then, the automatic measurement unit 13 detects a Nishikigoi from the photographic image for measurement after projective transformation, and estimates the body length of the Nishikigoi from the size of the chess board 35.

Specifically, for example, as shown in FIG. 9, when the distance from the water surface to the carp is 50 mm, the water depth is 150 mm, and the distance between the center of the chessboard 35 and the automatic Nishikigoi measuring device 10 is 800 mm, the automatic Nishikigoi measuring device 10 The distance from to Nishikigoi is
800mm-150mm+50mm=700mm
It is.

Next, the automatic measuring unit 13 measures the length of two squares of the chessboard 35 on the photographed image for measurement and the body length of the Nishikigoi on the photographed image for measurement, and measures the length of the two squares on the photographed image for measurement. When the length of the Nishikigoi is 58 pixels and the body length of the Nishikigoi on the photographed image for measurement is 360 pixels, the body length of the Nishikigoi is calculated by the following formula and becomes 315 mm.
700mm x 360 pixels/(800mm x 58 pixels/58mm)
=700mm x 360 pixels/800 pixels = 315mm

Then, the Nishikigoi automatic measuring device 10 acquires a photographic image for measurement, performs projective transformation processing, and estimates the body length of the Nishikigoi every second, and stores the respective determined values. Then, if four or more of the five consecutive measurement images do not deviate by 20% or more from the average of the five images, the measurement is determined to be successful. Therefore, the measurement can be completed in 5 seconds at the shortest, but if the measurement fails, it will time out in 10 seconds and prompt you to take the measurement again.

Next, the Nishikigoi evaluation device 20 will be explained. The Nishikigoi evaluation device 20 extracts feature amounts from the photographed image after color correction output from the Nishikigoi automatic measurement device 10, evaluates the Nishikigoi based on the feature amounts, and outputs the evaluation result to the Nishikigoi automatic measurement device 10. do. This makes it possible to objectively evaluate Nishikigoi, which was conventionally subjective.

Specifically, the Nishikigoi evaluation device 20 includes an image acquisition section 21, a feature amount acquisition section 22, an evaluation section 23, and an output section 24, as shown in FIG.

The image acquisition unit 21 acquires a photographed image of a Nishikigoi to be evaluated. Specifically, the image acquisition section 21 acquires a photographed image photographed by the photographing section 11 of the Nishikigoi automatic measuring device 10. The image acquisition unit 21 acquires, for example, an image of a Nishikigoi viewed from above (dorsal fin side). In this embodiment, the dorsal fin side of the Nishikigoi is the top, the ventral side is the bottom, the head side is the front, the tail side is the back, the right eye side is the right, and the left eye side is the left.

The feature amount acquisition unit 22 extracts and acquires feature amounts from the photographed image of the Nishikigoi to be evaluated, which is acquired by the image acquisition unit 21. In this embodiment, information on the body shape of the Nishikigoi, information on the color of the Nishikigoi, information on the pattern of the Nishikigoi, and information on the finish are acquired as the feature amounts.

The feature amount acquisition unit 22 performs contour extraction processing on the captured image acquired by the image acquisition unit 21. The contour extraction process is a process for extracting the contour of a Nishikigoi. Existing image processing such as edge extraction processing can be used for the contour extraction processing. The contour information of the Nishikigoi extracted by the contour extraction process is used when extracting the feature amount described below.

The feature acquisition unit 22 acquires information on the body shape of the Nishikigoi based on the contour information of the Nishikigoi extracted by the contour extraction process as described above, and also acquires color information based on the photographed image of the Nishikigoi after color correction. , obtain pattern information and finish information.

First, information on the body shape of Nishikigoi will be explained.

When evaluating Nishikigoi, the following points are taken into consideration regarding the body shape of Nishikigoi.
・The tail tube must be balanced in proportion to each part of the body, neither too thin nor too thick. ・The size of the head and the entire body must not be unbalanced. ・The size of each fin relative to the body must be Balanced and symmetrical - Arched with the back muscles and spine as the center line, symmetrical and without bending - Smooth line from the head to the caudal fin

Therefore, the feature amount acquisition unit 22 of this embodiment obtains a to f shown in FIG. 10 based on the contour information of the Nishikigoi. a is the width of the head and is the distance between the inside of the left eye and the inside of the right eye. Further, b is the shoulder width (width at the widest point of the body), and is the distance between the rearmost end of the left pectoral fin and the rearmost end of the right pectoral fin. c is the degree of protrusion of the abdomen, and is the distance between the left and right ends of the front end of the dorsal fin. d is the thickness of the caudal fin, and is the distance between the left and right ends at the front end of the caudal fin. e is the width of the pectoral fin, and is the distance from the base of the pectoral fin to the end in the left-right direction. f is the vertical width of the pectoral fin, and is the length from the front end to the rear end of the pectoral fin.

Note that the positions of the left eye and right eye, the position of the pectoral fin, the position of the dorsal fin, and the position of the caudal fin are detected by using existing image processing and pattern recognition.

Then, the feature amount acquisition unit 22 calculates the following formula to obtain the ratios Wa to Wf.
Wa=a/b, Wb=b/c, Wc=c/d, Wd=a/d, We=b/e,
Wf=e/f

Further, the feature amount acquisition unit 22 obtains g to k shown in FIG. 10 based on the contour information of the Nishikigoi. g is the total length of the body from the tip of the snout to the rear end of the caudal fin. h is the length of the head from the tip of the snout to the rearmost end of the gill cover. i is the length of the back from the most posterior end of the gill cover to the anterior end of the dorsal fin. j is the length of the body from the anterior end of the dorsal fin to the anterior end of the caudal fin. k is the length of the caudal fin from the anterior end of the caudal fin to the posterior end of the caudal fin.

Although the gill cover is not shown in FIG. 10, its position can be detected by extracting the gill outline. In addition, if it is difficult to detect a Nishikigoi with only an image taken from above, the image acquisition unit 21 further acquires an image taken from the left side or an image taken from the right side, and the feature amount acquisition unit 22 The position of the gill cover may be detected from the image.

Then, the feature amount acquisition unit 22 calculates the following formula to obtain the ratios Lh to Lf.
Lh=g/h, Li=g/i, Lj=g/j, Lk=g/k, Lf=g/f,

The feature amount acquisition unit 22 calculates the above-mentioned 11 ratios of Wa to Wf and Lh to Lf as information on the body shape of the Nishikigoi.

Further, the feature amount acquisition unit 22 obtains the spine line of the Nishikigoi from the Nishikigoi contour information. Specifically, the line from the front end of the dorsal fin to the front end of the caudal fin, passing through the dorsal fin, is determined as the line of the spine. For example, a straight line may be used to connect the rear end of the dorsal fin to the front end of the caudal fin.

Next, the feature amount acquisition unit 22 acquires information on the presence or absence of defects in the Nishikigoi as information on the body shape of the Nishikigoi.

Specifically, the feature amount acquisition unit 22 checks whether there are a total of four mustaches, two large and two small, based on the contour information of the Nishikigoi. Further, the feature amount acquisition unit 22 determines the positions of the left and right eyes from the image of the Nishikigoi, and checks whether the positions of the eyes are symmetrical. Furthermore, the feature amount acquisition unit 22 detects whether or not there is a warp in the gill cover based on the contour information of the Nishikigoi. Furthermore, the feature acquisition unit 22 checks whether the pectoral fin has a defect, twist, or deformation based on the outline information of the Nishikigoi.

Further, the feature amount acquisition unit 22 extracts scales from the Nishikigoi image and checks whether the scales are arranged in an irregular manner or not. Specifically, the feature amount acquisition unit 22 finds the center of each scale, and finds a straight line connecting the centers of scales adjacent in the front-back direction. Then, the feature amount acquisition unit 22 confirms the irregularity of the scales by detecting that the straight line is not smooth and includes irregularities.

In addition, the feature amount acquisition unit 22 obtains information on the presence or absence of keloids by confirming that there is no scale pattern within a preset area or more among the scale patterns extracted from the Nishikigoi image. get.

As described above, the feature acquisition unit 22 obtains information on the presence or absence of defects in the Nishikigoi, including mustache information, left and right eye symmetry information, gill cover warp information, pectoral fin defect information, and scale information. Obtain alignment information and keloid information.

That is, the feature amount acquisition unit 22 acquires the above-mentioned 11 ratios, the line of the spine, and 6 pieces of defect information as information on the body shape of the Nishikigoi.

Note that as the information on the spine line, for example, when the spine line is plotted in a preset coordinate system, the coordinates of a plurality of points on the line are acquired. Further, regarding the six pieces of defect information, for example, a numerical value of "1" is acquired as information indicating that there is a defect, and a numerical value of "0" is acquired as information that there is no defect.

Next, information on the color of Nishikigoi will be explained. In this embodiment, color information when the type of Nishikigoi is "red and white" will be explained.

When evaluating Nishikigoi (red and white), the following points are taken into consideration regarding the color information of Nishikigoi and the white background (=ground skin).
・Deep, thick, milky white color with a sense of transparency ・No yellowing (looks pink) on the ground ・No yellowing ・No stains etc.

Therefore, the feature amount acquisition unit 22 of this embodiment obtains the frequency distribution of the color of each pixel constituting the photographed image of the Nishikigoi, as shown in FIG. 11.

Then, the feature amount acquisition unit 22 determines the frequency of milky white pixels based on the color frequency distribution and acquires the frequency as white background information. Note that the RGB values of milky white are set in advance. The white background of Nishikigoi (Kohaku) is preferably milky white.

Further, the feature amount acquisition unit 22 determines the frequency of pink pixels corresponding to the color hi and acquires it as information about the color hi, based on the color frequency distribution. Note that it is assumed that the RGB values of pink corresponding to the yellow color are set in advance. The white background of Nishikigoi (red and white) is preferably milky white, and it is preferable that no pink color appears.

Further, the feature amount acquisition unit 22 obtains the frequency of yellow pixels corresponding to yellowing based on the color frequency distribution and acquires the frequency as yellowing information. Note that it is assumed that the RGB values of yellow corresponding to yellowing are set in advance. It is preferable that there is no yellowing.

Further, the feature amount acquisition unit 22 obtains information about the stain by determining the frequencies of brown and black pixels corresponding to the stain based on the color frequency distribution. Note that it is assumed that the brownish and blackish RGB values corresponding to the stain are set in advance. It is preferable that there are no stains.

Next, regarding information on the color of Nishikigoi, regarding the Hiban (red part), the following points are taken into consideration.
・The color of the hiban from the head to the tail tube must be uniform. ・The red color must be clear, and the hue (persimmon and pink colors are said to be bright and good) is good. ・It is not the chromaticity that is important. Checking the thickness of the Hiban (Neri-komi, that is, if you can't see the scales, it is considered thick)
・The marbling of the scarlet board is well-cut, and the edges of the head scarlet and the edges and marbling of the scales are not blurred.・There are no color shading or stains that stand out in the scarlet board.

Therefore, the feature amount acquisition unit 22 of this embodiment extracts the range of the scarlet board from the image of the Nishikigoi, and calculates the frequency distribution of the color of each pixel in the range of the scarlet board. It is assumed that the range of RGB corresponding to the range of scarlet is set in advance.

Then, the feature amount acquisition unit 22 obtains information on the uniformity of the color of the scarlet by finding the variance or standard deviation of the frequency distribution of colors in the range of the scarlet. The smaller the dispersion or standard deviation, the more uniform the color, which is preferable.

Further, the feature amount acquisition unit 22 converts the RGB of each pixel in the range of Hiban, for example, into a signal of the L ^* a ^* b ^* color space, and calculates the saturation based on the values of a ^* and b ^* . Obtained as information on the clarity (shade) of the color of the scarlet board.

In addition, the feature amount acquisition unit 22 extracts the scale pattern included in the entire image of the area of the Hiban, detects the edges of the extracted scale pattern, obtains the edge amount, and uses the information as information about the thickness of the Hiban. get. The edge amount is the total number of pixels forming an edge. The smaller the amount of edges, the thicker the Hiban, which is preferable.

Further, the feature amount acquisition unit 22 obtains the sharpness (degree of blur) of the edge at the boundary between the area of the scarlet board and the area of the white background based on the image of the Nishikigoi, and acquires this as information on the blurring of the marbling and creases. . Note that an existing method can be used to calculate the edge sharpness (degree of blur). It is preferable that the edge sharpness is high because there is less blurring of the edges and edges.

Note that the color shading and the fact that no stains stand out in the scarlet board are acquired as the above-mentioned stain information.

That is, the feature acquisition unit 22 obtains, as color information of the Nishikigoi, information on the white background, information on the lower scarlet color, information on yellowing, information on stains, information on the uniformity of the color of the scarlet disc, and information on the sharpness of the color of the scarlet disc. (color) information, information on the thickness of the scarlet board, and information on the bleeding of the marbling and edges.

Next, information on the pattern of Nishikigoi will be explained. In this embodiment, pattern information when the type of Nishikigoi is "red and white" will be explained.

When evaluating Nishikigoi (red and white), the following points are taken into consideration regarding the information on the Nishikigoi's pattern.
・The shape, size, and number of markings on each part of the animal, from the head through the body to the tail, must be in harmony with the body. ・Hiban must be cut off at both the tip of the snout and the base of the tail.

Therefore, the feature amount acquisition unit 22 of this embodiment extracts the range of the Hiban from the image of the Nishikigoi.
Then, the feature amount acquisition unit 22 determines whether or not a part of the range of the scarlet board is included in the range of the tip of the nose.
Check whether the range of Hiban is approaching the base of the tail. Note that the range of the tip of the nose and the base of the tail are set using an existing image processing method based on preset conditions.

If a part of the range of the Hiban is not included in the range of the tip of the nose, and the range of the Hiban is not approaching the base of the tail, the feature value acquisition unit 22 determines that the range of the Hiban is "good" as the coverage information of the Hiban. If part of the range of Hiban is included in the range of the tip of the nose, or if the range of Hiban is approaching the base of the tail, "Bad" is displayed as the information on Hiban. Get information about.

Furthermore, the feature acquisition unit 22 divides the image of the Nishikigoi into three parts: the head, the body, and the tail, and specifies the range of the Hiban included in each part. The head, body, and tail may be divided according to preset conditions.

Then, the feature amount acquisition unit 22 calculates the ratio of the area of the range of the scarlet board included in each part to the total area of each part, and acquires the ratio of each part as information on the balance of the scarlet board. Note that when the ratio of Hiban in each part is 50% to 90%, it is determined that the balance of Hiban is good, and in the case of other ratios, it is determined that the balance of Hiban is poor.

Further, the feature amount acquisition unit 22 acquires the number of scarlet boards included in each part and the circumferential length of each scarlet board as information on the balance of the scarlet boards.

That is, the feature amount acquisition unit 22 acquires information on the coverage of the hi-ban and information on the balance of the hi-ban as information on the pattern of the Nishikigoi. In addition, regarding the performance information of the scarlet board, for example, when the performance information of the scarlet board is "good", a numerical value of "1" is obtained, and when the performance information of the scarlet board is "bad", a numerical value of "0" is obtained.

Next, information on the finish of Nishikigoi will be explained. In this embodiment, information on the finish when the type of Nishikigoi is "red and white" will be explained.

When evaluating Nishikigoi (Kohaku), the following points are taken into consideration regarding the finish information of Nishikigoi.
・The color must be clear and vivid, with gloss and shine, and the presence of cloudiness. ・The scalp must be transparent, glossy, glossy, and smooth, and there is no bloodshot. ・The contrast between the scarlet color and the scalp must be clear. (must be a bright color)

Therefore, the feature amount acquisition unit 22 determines the contrast, brightness, and saturation based on the image of the Nishikigoi and acquires them as finished information. Regarding the saturation, the saturation of the Hiban area was determined as information on the color of the Nishikigoi, but the saturation of the entire Nishikigoi image was determined as information on the finish.

The above is the description of the feature amounts acquired by the feature amount acquisition unit 22 of this embodiment.

Next, the evaluation section 23 will be explained.

The evaluation unit 23 has a learned model obtained by performing machine learning on the relationship between the above-described feature amounts of a plurality of Nishikigoi and the evaluation results of each Nishikigoi. More specifically, the evaluation unit 23 of the present embodiment includes a trained model for body shape evaluation for evaluating the body shape of the Nishikigoi to be evaluated, a trained model for quality evaluation for evaluating the quality of the Nishikigoi, It has four trained models: a trained model for pattern evaluation to evaluate the pattern of Nishikigoi and a trained model for finish evaluation to evaluate the finish of Nishikigoi.

As for the trained models for body shape evaluation, for example, the above-mentioned body shapes of the 10 Nishikigoi that won world champions (overall overall championships) at past International Nishikigoi Shows and All Japan Nishikigoi Shows sponsored by All Japan Airinkai and All Japan Promotion Association. Machine learning is performed on the relationship between the information on the body shape and the evaluation results (scores) regarding the body shape, and a learned model is generated as a reference. Further, for the learned model serving as the reference, a learned model for body shape evaluation is obtained by performing machine learning on the relationship between the above-mentioned body shape information of any plurality of Nishikigoi and the evaluation results (scores) regarding the body shape. The above-described plurality of Nishikigoi include Nishikigoi with various evaluation results, such as those with good evaluation results and those with poor evaluation results.

As for the trained model for quality evaluation, machine learning is performed on the relationship between the above-mentioned color information of the 10 world champion Nishikigoi and the evaluation results (scores) regarding the color, and a trained model serving as a reference is generated. Then, for the trained model serving as the standard, a trained model for quality evaluation is obtained by performing machine learning on the relationship between the above-mentioned color information of any plurality of Nishikigoi and the evaluation result (score) regarding that color.

Regarding the trained model for pattern evaluation, machine learning is performed on the relationship between the above-mentioned pattern information of the 10 world champion Nishikigoi and the evaluation results (scores) regarding the pattern, and a learned model serving as a reference is generated. Further, for the learned model serving as the reference, a learned model for pattern evaluation is obtained by performing machine learning on the relationship between the above-mentioned pattern information of arbitrary Nishikigoi and the evaluation results (scores) regarding the pattern.

The learned model for finish evaluation performs machine learning on the relationship between the above-mentioned finish information of the 10 world champion Nishikigoi and the evaluation results (scores) regarding the finish, and generates a learned model that serves as a reference. Then, for the learned model serving as the standard, a learned model for finish evaluation is obtained by performing machine learning on the relationship between the above-mentioned finish information of any plurality of Nishikigoi and the evaluation results (scores) regarding the finish.

Note that known methods can be used as the machine learning method, such as deep neural network (DNN), convolutional neural network (CNN), recurrent neural network (RNN), and denoising stack autoencoder (DSA). can be used.

The evaluation unit 23 evaluates the Nishikigoi to be evaluated using four trained models: a trained model for body evaluation, a trained model for quality evaluation, a trained model for pattern evaluation, and a trained model for finish evaluation. .

Specifically, the evaluation unit 23 determines the body shape of the Nishikigoi to be evaluated by inputting the body shape information of the Nishikigoi to be evaluated, which was acquired by the feature value acquisition unit 22, into the trained model for body shape evaluation. Evaluate and obtain evaluation results (scores).

In addition, the evaluation unit 23 evaluates the quality of the Nishikigoi to be evaluated by inputting the color information of the Nishikigoi to be evaluated acquired by the feature amount acquisition unit 22 to the trained model for quality evaluation, Obtain evaluation results (scores).

In addition, the evaluation unit 23 evaluates the pattern of the Nishikigoi to be evaluated by inputting information on the pattern of the Nishikigoi to be evaluated acquired by the feature amount acquisition unit 22 to the trained model for pattern evaluation, Obtain evaluation results (scores).

In addition, the evaluation unit 23 evaluates the finish of the Nishikigoi to be evaluated by inputting the finish information of the Nishikigoi to be evaluated acquired by the feature value acquisition unit 22 to the trained model for finish evaluation, Obtain evaluation results (scores).

As described above, the evaluation unit 23 obtains the body shape score, color score, pattern score, and finish score of the Nishikigoi to be evaluated.

Furthermore, the evaluation unit 23 obtains a comprehensive evaluation result (total score) of the Nishikigoi to be evaluated by multiplying each of the above-mentioned four scores by preset weighting and adding them.

The weighting can be set arbitrarily; for example, the body type score is weighted k1, the quality score is weighted k2, the pattern score is weighted k3, the finish score is weighted k4, and the body shape score is weighted s1. , where the quality score is s2, the pattern score is s3, and the finish score is s4, the total score sa is calculated by the following formula (1).
sa=k1×s1+k2×s2+k3×s3+k4×s4...(1)

In this embodiment, the weights k1 to k4 are changed depending on the growth stage of the Nishikigoi to be evaluated.

For example, if the Nishikigoi to be evaluated is at the juvenile stage ((Parts 12 to 25) 1 cm to 25 cm), the weighting is set to 25% for body shape, 25% for quality, 25% for pattern, and 25% for finish. . In addition, if the Nishikigoi to be evaluated is a young carp ((parts 30 to 40) 25.1cm to 40cm), the weighting is 35% for body type, 30% for quality, 20% for pattern, and 15% for finish. is set to In addition, if the Nishikigoi to be evaluated is an adult fish ((Parts 45 to 55) 40.1cm to 55cm), the weighting will be 40% for body type, 30% for quality, 20% for pattern, and 10% for finish. Set. In addition, if the Nishikigoi to be evaluated is at the stage of a magnificent fish ((Parts 60 to 70) 55.cm to 70cm), the weighting is 40% for body type, 30% for quality, 20% for pattern, and 10% for finish. is set to In addition, if the Nishikigoi to be evaluated is at the stage of giant carp ((Parts 75 to 90) 70.1cm to 90cm), the weighting will be 50% for body type, 25% for quality, 15% for pattern, and 10% for finish. Set.

The evaluation unit 23 receives the body length of the Nishikigoi measured by the Nishikigoi automatic measuring device 10, determines the growth stage of the Nishikigoi based on the received body length of the Nishikigoi, and sets weighting according to the determined growth stage. , calculate the above-mentioned total score sa.

In the present embodiment, the evaluation unit 23 has four learned models, but the present invention is not limited to this, and the four learned models may be connected to the Nishikigoi evaluation device 20 via a communication network. The information may be stored in a server device, etc., and may be used for evaluation.

Further, the Nishikigoi evaluation device 20 of this embodiment is provided with a trained model generation section, and the trained model generation section generates a trained model for body evaluation, a trained model for quality evaluation, a trained model for pattern evaluation, and a trained model for pattern evaluation. Four trained models for evaluation may be generated. In that case, the image acquisition section 21 may acquire the photographed image of the Nishikigoi for machine learning described above, and the feature amount acquisition section 22 may acquire the feature amount of the acquired photographed image. Further, the evaluation results (scores) of each Nishikigoi may be set and input by the user using a predetermined input device (not shown) connected to the Nishikigoi evaluation device 20.

The output unit 24 outputs the evaluation result of the Nishikigoi to be evaluated, which has been evaluated by the evaluation unit 23. The output unit 24 of this embodiment outputs the body type score s1, quality score s2, pattern score s3, finish score s4, and total score sa of the Nishikigoi to be evaluated to the Nishikigoi automatic measuring device 10.

Specifically, the output unit 24 of this embodiment displays the body shape score s1, quality score s2, pattern score s3, finishing score s4, and total score sa of the Nishikigoi to be evaluated on the Nishikigoi automatic measuring device 10. section 15. As a display method, for example, as shown in FIG. 12, body shape score s1, quality score s2, pattern score s3, and finish score s4 are displayed as a radar chart, and the total score sa is displayed in text.

The Nishikigoi evaluation device 20 is composed of a computer, which includes a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), semiconductor memory such as ROM (Read Only Memory) and RAM (Random Access Memory), storage such as a hard disk, and communication. It is equipped with I/F (Interface), etc.

Furthermore, a Nishikigoi evaluation program is installed in the semiconductor memory or hard disk of the Nishikigoi evaluation device 20. When the CPU or GPU executes this Nishikigoi evaluation program, the image acquisition section 21, feature amount acquisition section 22, evaluation section 23, and output section 24 described above function.

Note that the hardware configuration of the Nishikigoi evaluation device 20 is not limited to the configuration described above.

In addition, in this embodiment, all of the functions of the image acquisition section 21, feature value acquisition section 22, evaluation section 23, and output section 24 described above are executed by the Nishikigoi evaluation program, but the present invention is not limited to this. Alternatively, all functions may be configured from hardware such as an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or other electrical circuits.

Next, the process flow of the Nishikigoi automatic measurement and evaluation system 1 of this embodiment will be explained with reference to the flowchart shown in FIG. 13.

First, the photographing section 11 of the automatic Nishikigoi measuring device 10 is activated, and photographing frames 15a and 15b are displayed on the display section 15 (S10).

Next, the color pallet 30 with scales is placed near the Nishikigoi to be measured (S12), so that both the Nishikigoi and the color pallet 30 with scales are within the photographing range of the photographing unit 11.

Subsequently, the user adjusts the position and angle of the automatic Nishikigoi measuring device 10 so that the Nishikigoi in the preview image displayed on the display unit 15 falls within the photographing frames 15a and 15b (S14). Then, when the Nishikigoi in the preview image is placed within the photographing frames 15a and 15b by the photographing unit 11 (S16, YES), photographing becomes possible (S18), and the photographing button 14c is tapped to photograph the photographed image ( S20).

Next, the measurement target portion recognition section 12 recognizes the head and caudal fin of the Nishikigoi in the photographed image (S22), and the automatic measurement section 13 automatically measures the body length of the Nishikigoi (S24). Furthermore, the color correction unit 14 performs color correction on the photographed image (S26).

Subsequently, the image acquisition unit 21 of the Nishikigoi evaluation device 20 acquires a photographed image after color correction (S28).

Then, the feature amount acquisition unit 22 obtains feature amounts of body shape information, color information, pattern information, and finish information of the Nishikigoi based on the color-corrected photographed image (S30).

Then, the respective feature quantities are input to the four learned models possessed by the evaluation unit 23, and the body shape score, quality score, pattern score, finish score, and overall score of the Nishikigoi to be evaluated are determined ( S32).

Then, the output unit 24 outputs the body type score, quality score, pattern score, and finish score of the Nishikigoi to be evaluated to the Nishikigoi automatic measurement device 10 and displays them on the display unit 15 as a radar chart, and also 15 to display the total score in text. (S34).

In the Nishikigoi evaluation device 20 of the above embodiment, the information on the presence or absence of defects in Nishikigoi includes mustache information, left and right eye symmetry information, gill cover warp information, pectoral fin defect information, scale alignment information, and Information such as keloids is obtained and evaluated as information on the Nishikigoi's body shape, but this is not limited to this. Information such as defects and defects of the Nishikigoi can be evaluated independently and multiplied by this as a coefficient. Alternatively, the total score sa may be calculated. That is, the total score sa may be calculated by multiplying coefficients m1 to m4 based on the evaluation results of information such as defects and defects of the Nishikigoi, as shown in the following equation (2).
sa=k1×s1×m1+k2×s2×m2+k3×s3×m3+k4×s4×m4...(2)

Furthermore, for example, in Nishikigoi competitions held by different organizers, the weights given to the four evaluation items of Nishikigoi body shape, Nishikigoi color, Nishikigoi pattern, and finish are different. That is, there are some Nishikigoi fairs where evaluation of the Nishikigoi's body shape is given more weight than other evaluation items, and there are also Nishikigoi fairs where evaluation of the Nishikigoi's pattern is given more weight than other evaluation items. Therefore, the total score may be calculated by multiplying each evaluation item by a different coefficient for each Nishikigoi fair. The total score sa may be calculated based on the following formula (3) using different coefficients n1 to n4 for each Nishikigoi fair.
sa=k1×s1×m1×n1+k2×s2×m2×n2+k3×s3×m3×n3+k4×s4×m4×n4...(3)

n1 to n4 are values arbitrarily set by the organizer of the Nishikigoi fair.

Further, in the above embodiments, the organism to be measured is a colored carp, but the organism to be measured in the present invention is not limited to a colored carp, but may be an ornamental fish such as a goldfish, or an ornamental plant such as a bonsai. When measuring an ornamental plant, the photographing unit 11 displays a photographing frame representing the external shape of the plant. Then, the measurement target portion recognition unit 12 recognizes the root and the highest tip of the plant, and the automatic measurement unit 13 measures the distance between the root and the highest tip of the plant (the height of the plant). Note that the part to be measured is not limited to the above, and may be any part that is characteristic of the plant.

Regarding the present invention, the following additional notes are further disclosed.
(Additional note)

In the automatic biological measurement device of the present invention, the biological object to be measured is an ornamental fish, the measurement target part recognition unit recognizes the head and caudal fin of the ornamental fish as the measurement target part, and the automatic measurement unit recognizes the head and caudal fin of the ornamental fish as the measurement target part. The body length of an ornamental fish can be automatically measured based on the head and caudal fin of the ornamental fish.

In the automatic biological measurement device of the present invention, the scale information is included in the photographed image, and the automatic measurement unit can automatically measure the living organism based on the scale information in the photographed image.

In the biological automatic measurement device of the present invention, information on color samples of preset colors is included in the photographed image, and the color of the photographed image is corrected based on the information on the color sample in the photographed image. A correction section can be provided.

In the biological automatic measuring device of the present invention, the captured image can be an image of an underwater creature.

The automatic biological measuring device of the present invention includes a photographing section that photographs a photographed image, and the photographing section can display a photographing frame representing the external shape of the living organism.

In the automatic biological measuring device of the present invention, the photographing unit can recognize whether or not a living thing is within the photographing frame, and can permit photographing of a photographed image only when the living thing is within the photographing frame. .

In the biological automatic measuring device of the present invention, the photographing frame can represent the external shape of the ornamental fish.

1 Nishikigoi automatic measurement evaluation system 10 Nishikigoi automatic measurement device 11 Photographing section 12 Measurement target part recognition section 13 Automatic measurement section 14 Color correction section 14c Shooting button 15 Display sections 15a, 15b Photographing frame 20 Nishikigoi evaluation device 21 Image acquisition section 22 Feature amount Acquisition unit 23 Evaluation unit 24 Output unit 30 Color palette with scale 31 Color chart 31a Rectangular patterns 32a to 32d Scale unit 32e Scale pattern 32f Interval 33 First sheet 34 Second sheet

Claims (10)

a measurement target part recognition unit that recognizes a measurement target part of the creature in a captured image of the creature to be measured;
An automatic biological measurement device comprising: an automatic measurement section that automatically measures the living organism based on the measurement target portion recognized by the measurement target portion recognition section. The organism to be measured is an ornamental fish,
The measurement target part recognition unit recognizes the head and caudal fin of the ornamental fish as the measurement target part,
The automatic biological measuring device according to claim 1, wherein the automatic measuring unit automatically measures the body length of the ornamental fish based on the recognized head and caudal fin of the ornamental fish. Scale information is included in the captured image,
The automatic living organism measuring device according to claim 1, wherein the automatic measuring unit automatically measures the living organism based on scale information in the photographed image. The photographed image includes information on a color sample of a preset color,
The automatic biological measurement device according to claim 1, further comprising a color correction section that corrects the color of the photographed image based on information of the color sample in the photographed image. 5. The automatic biological measuring device according to claim 4, wherein the captured image is an image of the living organism underwater. comprising a photographing unit that photographs the photographed image;
The automatic biological measuring device according to claim 1, wherein the photographing section displays a photographing frame representing the external shape of the living organism. The living thing according to claim 6, wherein the photographing unit recognizes whether or not the living thing is within the photographing frame, and permits photographing of the photographed image only when the living thing is within the photographing frame. Automatic measuring device. 7. The automatic biological measuring device according to claim 6, wherein the photographic frame represents the external shape of the ornamental fish. Recognizing the measurement target part of the creature in a photographed image of the creature to be measured,
An automatic biological measurement method for automatically measuring the living organism based on the recognized measurement target part. a step of recognizing a part of the organism to be measured in a captured image of the organism to be measured;
A biological automatic measurement program that causes a computer to execute a step of automatically measuring the biological body based on the recognized measurement target part.